复杂环境下基于视觉的移动机器人导航控制和虚拟现实关键技术研究

工业4.0时代的到来，对机器人的自主性、智能性提出了更高的要求。传统的导航方法已无法满足移动机器人在未知的、动静态混合的自然环境中自主作业的需要，一种有效的、系统的、针对复杂环境的新一代移动机器人自主导航方法具有重大的研究价值。本项目将针对上述问题，融合控制理论、传感器技术、计算机视觉及人机交互等先进信息技术，提出一种以机器人视觉为核心的导航控制框架，构建包括视觉同步定位与建图、组合动态轨迹规划和视觉伺服控制在内的完整导航理论体系，并开发相应的虚拟现实仿真平台。着力研究真实环境中光照变化、动态障碍物等对视觉导航带来的技术难题，提出一套精度高、鲁棒性强的系统导航解决方案。最终建立基于视觉的自主导航移动机器人实验样机，并在真实环境中开展测试。复杂环境下的移动机器人导航控制理论研究，为实际应用中服务机器人或无人驾驶车辆的自主导航提供可行方法，大大提高机器人在非结构化环境下的自主能力。

With the arrival of Industrie 4.0, robots are required to work in more challenging environment and perform demanding tasks. Since traditional navigation methods is no longer sufficient in this condition, there is an urgent for the next generation of navigation algorithm to guide mobile robots through complex unstructured environment safely, efficiently and robustly. To solve the problem, this project inosculates the advanced techniques in control theory, sensor technology, computer vision and human-robot interaction, and proposes a vision-based navigation framework in complex environment. A complete theory system of mobile robot vision-based navigation will be constructed, including visual simultaneous localization and mapping (V-SLAM), combined trajectory planning with dynamic obstacles considered and self-adapting visual servo control. A simulation software equipped with virtual reality techniques will be developed. This project aims at the deep-going research on the technological means of eliminating the disturbance of varying illumination, adapting to the dynamic complex environment and improving the performance prominently. This project tries to find the innovative solutions to the dilemma of robustness, accuracy and efficiency in the existing algorithms. A mobile robot prototype will be built to conduct the test and evaluate the adaptability and reliability of the proposed methods and algorithms in dynamic real environments. The study on vision-based navigation methods in complex environment offers a feasible way to autonomous navigation of service robots or driverless cars and improves autonomy of robots in a large extent.